outlook-hack
Your agent reads Outlook email all day. Drafts replies for you. Won't send a single one. 90 days per browser tap.
Why use this skill?
Use the Outlook Hack skill to safely read, index, and draft emails in your Outlook mailbox using OpenClaw AI. Privacy-focused and read-only.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/globalcaos/outlook-hackWhat This Skill Does
The outlook-hack skill provides OpenClaw agents with deep, read-only access to Microsoft Outlook mailboxes. By leveraging an ingenious bypass of modern PoP token restrictions, it allows your agent to interact with your professional communications safely. The agent can ingest entire folder contents, index attachments, and generate intelligent summaries of your daily correspondence. Crucially, the system is hard-coded to prevent the agent from ever sending an email, ensuring you never experience accidental or unwanted outgoing communication.
Installation
Installation is a two-step process requiring the extraction of a Microsoft Teams MSAL refresh token. First, open the Microsoft Teams web interface in your browser and use the browser console to extract the refresh token from your localStorage. Save this token along with your tenant ID and client details into the ~/.openclaw/credentials/outlook-msal.json configuration file. Once configured, verify the connection using the provided test script. Finally, use the clawhub install openclaw/skills/skills/globalcaos/outlook-hack command to complete the agent setup. This process effectively grants the agent legitimate access to the Microsoft Graph API using the existing permissions assigned to your Teams client.
Use Cases
This skill is perfect for professionals overwhelmed by inbox volume. It enables users to offload the burden of sorting through hundreds of daily messages. Use it to automatically summarize long email chains, identify urgent requests that require your manual attention, or track project attachments across months of history. It serves as an intelligent filter between your email server and your workspace.
Example Prompts
- "Summarize all unread emails from my team today and highlight any requests that need a response."
- "Search for the latest budget report attachment from last week and summarize the core findings."
- "Draft a response to the recent client inquiry in my inbox thanking them for the update and confirming that I will review the file by tomorrow."
Tips & Limitations
This skill is strictly read-only for emails; drafting is supported, but actual transmission is physically disabled by design. Because it utilizes a refresh token, you may need to re-extract the token every 90 days as the session expires. Ensure your ~/.openclaw/ directory remains secure, as these credentials provide read access to your email data. Use the provided command-line utilities for efficient bulk fetching to avoid performance degradation during large indexing jobs.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-globalcaos-outlook-hack": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: network-access, file-write, file-read, data-collection, external-api
Related Skills
jarvis-voice
Turn your AI into JARVIS. Voice, wit, and personality — the complete package. Humor cranked to maximum.
shell-security-ultimate
Classify every shell command as SAFE, WARN, or CRIT before your agent runs it.
memory-pioneer
Benchmark your agent's memory. Contribute anonymized scores to open research. Citizen science for AI memory.
subagent-overseer
Monitor sub-agent health and progress via a pull-based bash daemon. Use when spawning sub-agents that need progress tracking, staleness detection, and automatic status reporting. Replaces manual heartbeat polling with a deterministic status file the agent reads every 3 minutes. Zero AI tokens for monitoring — pure OS-level process checks and filesystem diffs.
model-router
Automatic LLM model selection for sub-agent tasks. Classifies tasks by complexity and type, then routes to the optimal model (cost vs capability). Use when spawning sub-agents, choosing models for cron jobs, or deciding which model to use for any task. Eliminates manual model specification by providing a decision tree and optional cheap-model classifier for ambiguous cases.